A satellite-linked tag for the long-term monitoring of diving behavior in large whales

• Published in: Animal Biotelemetry
• Main Authors: Daniel M. Palacios, Ladd M. Irvine, Barbara A. Lagerquist, James A. Fahlbusch, John Calambokidis, Stanley M. Tomkiewicz, Bruce R. Mate
• Format: Article in Journal/Newspaper
• Language: English
• Published: BMC 2022
• Subjects: Argos satellite telemetry; Data recovery; Bio-logging; Smart tags; Accelerometry; Adaptive algorithm; Ecology; QH540-549.5; Animal biochemistry; QP501-801; Rorqual; Balaenoptera musculus; Megaptera novaeangliae
• Online Access: https://doi.org/10.1186/s40317-022-00297-9; https://doaj.org/article/1589c259700540a0a4d1fb3847835eac

Abstract Despite spending most time underwater, the technology in use to track whales over large geographic ranges via satellite has been largely limited to locational data, with most applications focusing on characterizing their horizontal movements. We describe the development of the RDW tag, a new Argos-based satellite telemetry device that incorporates sensors for monitoring the movements and dive behavior of large whales over several months without requiring recovery. Based on an implantable design, the tag features a saltwater conductivity switch, a tri-axial accelerometer, and an optional pressure transducer, along with onboard software for data processing and detection of behavioral events or activities of interest for transmission. We configured the software to detect dives and create per-dive summaries describing behavioral events associated with feeding activities in rorqual whales. We conducted a validation by proxy of the dive summary and event detection algorithms using field data from a medium-duration archival tag. We also conducted a simulation exercise to examine how the expected data recovery would vary under different dive behavior scenarios and compared those results to empirical values from field deployments of the RDW tag on blue (Balaenoptera musculus) and humpback (Megaptera novaeangliae) whales. The dive summary algorithm accurately reported dive depth and duration, while the accuracy of the lunge-feeding event detection algorithm was dependent on the precision of the accelerometer data that was used, with a predicted accuracy of 0.74 for correctly classifying feeding dives from 1/64-G precision data and 0.95 from 1-mG precision data. Simulated data recovery was lower with sparser transmission schedules, shorter mean dive durations, and lower rates of successfully received transmissions. Empirical data recovery was lower than expected from the simulation, suggesting the effect of additional factors, such as data gaps. By measuring key aspects of the per-dive behavior of large whales ...